Complexity matters in TV dialogue as language input. A corpus-based description of clausal and phrasal patterns

The spread of digital devices and new media has led to a proliferation of English-language audiovisual products and to a gradual shift in learning English as a second language (L2), which is increasingly occurring outside of the classroom. In fact, the extensive informal contact with English audiovisual dialogue as part of L2 learner-users’ everyday leisure activities has been found to be potentially conducive to incidental learning and language acquisition (Kusyk 2020; Sockett 2014). This calls for a description of English-language media, the major source of language input to which L2 viewers are exposed (Pavesi, Ghia 2020). The present study aims to investigate grammatical complexity in fictional TV dialogue by drawing on the Sydney Corpus of Television Dialogue (SydTV; Bednarek 2018a). A register-functional (RF) approach (Biber 1988) is adopted for the quantitative and qualitative assessments of clausal and phrasal complexity features in TV series, with a focus on finite/nonfinite subordinate clauses and noun phrase premodification. The data are first interpreted in relation to the diegetic and extradiegetic functions served by complexity features onscreen and are then compared to previous corpus-based findings regarding the grammatical complexity of spontaneous face-to-face conversations (Biber 2015; Biber et al. 2021). The results show that TV dialogue closely approximates casual conversation in terms of the main patterns of phrasal and clausal complexity. Register-specific functions emerge in accordance with the audience-oriented narrative dimension of telecinematic products, the striving for realism and the expression of characters’ stances. Such referential and communicative functions may increase the accessibility of TV narratives for L2 viewers, often in tandem with visuals in a multimodal fashion. Overall, by reproducing the complexity of conversational exchanges, TV dialogue qualifies as a rich, reliable source of input that learner-viewers can readily use as a model of spoken English.